Fig. 1

Overview of SeqOthello structure and query procedure. a An illustration of the SeqOthello indexing structure to support scalable k-mer searching in large-scale sequencing experiments. The bottom level of SeqOthello stores the occurrence maps of individual k-mers, encoded in three different formats and divided into disjoint buckets. The mapping between a k-mer and its occurrence map is achieved by a hierarchy of Othello structures in which the root Othello maps a k-mer to its bucket and the Othello in each bucket maps a k-mer to its occurrence map. b An example illustrating SeqOthello’s sequence query process and output. A sequence query is decomposed into its constituent k-mers. The query result can be either a k-mer hit map, recording each k-mer’s presence/absence along the query sequence, or k-mer hit ratios (i.e., the fraction of query k-mers present in each experiment)